Fuzzy systems and controllers - Definition and cognition

Since its definition by L.A. Zadeh [30] and the introduction of fuzzy controller principles by E.H. Mamdani [15] , fuzzy logic has become a standard tool for engineers.

The objectives of this article and the following one . By associating it with the articles [A 120] or by A. Kaufmann [20] and [R 7 428] by J.-P. Barrat, M. Barrat and Y. Lécluse [28] , it is intended to familiarize non-specialists, technicians and engineers with fuzzy systems, and fuzzy controllers in particular, in a description that is deliberately intended to be "pedagogical". It will therefore extend certain theoretical definitions proposed by A. Kaufmann and practical notions introduced by J.-P. Barrat, M. Barrat and Y. Lécluse. In this way, through the multiple extensions we propose, this article will provide automaticians with the scientific and/or empirical arguments they need to address the issues involved in choosing fuzzy logic, and to guide them in their choice of the most appropriate technical solution.

This topic is the subject of two articles, each divided into two parts.

1. The first part of [S 7 440] will be devoted to a review of the principles and definitions of multi-input, multi-output fuzzy systems and controllers (MEMS or MIMO, for Multi-inputs – Multi-Outputs). A detailed algorithm for hardware and software implementation will then be presented.

2. In the second part of [S 7 440], we will first deal with the synthesis of fuzzy controllers by presenting a set of rules of thumb and the most common structures for PID-type fuzzy controllers; then we will look at the tuning of fuzzy controllers by: either a qualitative tuning method, or a self-tuning method modelled on the Ziegler-Nichols approach defined for conventional controllers.

3. The first part of will describe a concrete experimental application on a chopper-devolver.

4. Finally, we'll conclude with with a summary of the main ideas developed here, some possible industrial solutions for the...